Statistical Learning Method 7-Support Vector Machine

Support Vector Machine: Discriminant Model
Type: Hard Interval, Soft Interval, Kernel Function
Idea: Perceptron is a linear classifier, and there can be infinitely many hyperplanes to separate linearly separable data sets. The support vector machine selects the most reliable one among them, and the basis for this reliable is: (Important Idea): 1. Under the premise of accurate classification, 2. Make the closest point to the hyperplane the longest distance from the hyperplane Far, that is to say, a point may be the closest point of many hyperplanes, but we compare the distance between this point and other hyperplanes, and choose the farthest hyperplane as the final hyperplane.
Learning process: the input space is mapped to the feature space, and the feature space is learned to obtain the output space.

Reference:
https://blog.csdn.net/Smile_mingm/article/details/108420581?spm=1001.2014.3001.5501